JP2004117368A - Acoustic sensor - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an acoustic sensor which conducts spectral analysis of an acoustic signal on a piece of hardware rapidly and precisely. <P>SOLUTION: An acoustic wave received by a wave receiving part 24 propagates through a rod-like resonance part 21 of a cantilever having a rod-like resonator 25 with a length resonating at each different specific frequency, and the resonator 25 of the resonance part 21 resonates. The resonance changes the capacitance of the capacitor comprising the tip of each resonator 25 and an electrode 3. The change in capacitance is converted into an electric signal by a detecting circuit 4, and the electric signal is integrated within an arbitrary set period to detect a time-varying change in a level of an acoustic wave on each frequency. <P>COPYRIGHT: (C)2004,JPO

Description

The present invention relates to an acoustic sensor for extracting characteristics of a sound signal in voice recognition processing, acoustic signal processing, and the like, and particularly to an acoustic sensor for detecting the intensity of a sound signal in each frequency band.

Conventionally, in a system for performing voice recognition, a microphone vibration receiving a voice signal is converted and amplified into an electrical signal by an amplifier, and then an analog signal is digitized by an A / D converter to obtain a voice digital signal. The voice digital signal is subjected to a fast Fourier transform by software on a computer to extract voice features. Such a speech recognition system is disclosed (for example, see Non-Patent Document 1).

In order to efficiently extract the features of the audio signal, it is necessary to calculate the acoustic spectrum within a time period in which the audio signal can be considered to be stationary. In the case of an audio signal, it is generally considered that the audio signal can be regarded as stationary within a time of 10 to 20 msec. Therefore, a signal processing such as a fast Fourier transform is executed by software on a computer for the audio digital signal included in the time period of 10 to 20 msec.

As described above, in the conventional voice recognition system, a voice signal including the entire instantaneous band is converted into an electric signal by the microphone, and A / D conversion is performed to analyze the spectrum of the electric signal to perform each frequency conversion. Is digitized, and the voice digital signal data is compared with data of a specific voice waveform to extract voice characteristics.

By the way, the auditory mechanism and the psychophysical properties of sound have been described in detail (for example, see Non-Patent Document 2). This document indicates that the scale of the pitch (pitch) of a sound heard by a human does not linearly correspond to the frequency as a physical quantity, but linearly corresponds to a scale called a mel scale. This mel scale indicates a psychological attribute (psychological scale) that represents the pitch of a sound as expressed in a scale, and is a scale that directly quantifies the frequency intervals called pitches that can be heard at equal intervals by humans. And the pitch of the sound at 1000 Hz and 40 phones is defined as 1000 mel. The 500-mel sound signal sounds as a 0.5-pitch sound, and the 2000-mel sound signal sounds as a 2-pitch sound. This mel scale can be approximated by the following equation (1) using a frequency f [Hz] as a physical quantity. FIG. 7 shows the relationship between the pitch [mel] and the frequency [Hz] in this approximate expression.
mel = (1000 / log2) log (f / 1000 + 1) (1)

In order to efficiently extract features of speech, it is common to convert a frequency band of an acoustic spectrum into such a mel scale. The conversion of the acoustic spectrum to the mel scale is usually performed by software on a computer, as in the analysis of the spectrum.

と し て In addition, as a technique for efficiently extracting features of speech, conversion of a frequency band of an acoustic spectrum to a bark scale is often performed. This bark scale is a measure corresponding to the loudness of a person's psychological sound (loudness), and indicates the frequency bandwidth over which humans can hear the sound (loudness is called the critical bandwidth). Sound within this critical bandwidth sounds the same at different frequencies. For example, when a large noise occurs in the critical bandwidth, the scale indicating a frequency band in which the noise and the signal sound cannot be distinguished by human hearing even though the signal sound has a different frequency from the noise. It is a bark scale.

In the field of audio signal processing, a critical bandwidth that is easy to handle on a computer is required, and the frequency axis of the acoustic spectrum is indicated by a bark scale that defines one critical band as 1 bark. FIG. 8 shows a numerical relationship between the critical bandwidth and the Bark scale. Also, these critical bandwidths and bark scales can be approximated by the following equations (2) and (3) using the frequency f [kHz] as a physical quantity.
Critical bandwidth: CB [Hz] = 25 + 75 (1 + 1.4 f ² ) ^0.69 (2)
Bark scale: B [Bark]
= 13 tan ^-1 (0.76f) +3.5 tan ^-1 (f / 7.5) ... (3)

By the way, it is known to use an engineering function model of the auditory peripheral system in the field of speech recognition, and the above-mentioned Non-Patent Document 2 provides a detailed explanation. In the engineering function model, the frequency spectrum analysis by the band filter group is used as the pre-processing. For example, in the pre-processing in the model of Seneff, which is one of the typical engineering function models, forty independent frequencies in the 130 to 6400 Hz frequency domain The frequency spectrum analysis is performed by a group of critical bandwidth filters having the specified channels. At this time, the frequency band of the acoustic spectrum is converted to Barks scale. In this model, the output of the model with respect to the input sound stimulus is obtained by computer simulation, and it is shown that the model matches well with the physiological data. Therefore, by using such an engineering function model, it is possible to improve the speech recognition rate in noise in automatic speech recognition.

The frequency spectrum analysis and the conversion of the acoustic spectrum to the mel scale by such a critical bandwidth filter group are usually executed by software on a computer, similarly to the spectrum analysis.
IEEE Signal Processing Magazine, Vol.13, No.5, pp.45-57 (1996) Supervised by Shunichi Amari, Seichi Nakagawa, Kiyohiro Kano, Yoichi Higashikura, "Neuroscience & Technology Series Speech / Hearing and Neural Network Models" (Ohmsha, 1992)

従 来 The conventional method of performing a fast Fourier transform process on a digital audio signal by software on a computer and analyzing the spectrum of the audio signal has a problem that the amount of calculation is enormous and the calculation load is large. Also, when a series of processes for converting a spectrum of an acoustic signal into a fast Fourier transform and converting it to a mel scale is performed by software on a computer, the amount of calculation is enormous and the calculation load is large. Furthermore, when a series of processes for analyzing the spectrum of an acoustic signal by a critical bandwidth filter group and converting it to bark scale is performed by software on a computer, the amount of calculation is enormous and the calculation load is large.

In addition, in the conventional method, there is no problem for a voice such as a vowel in which the acoustic spectrum does not change with time, but a sound composed of a combination of a consonant and a vowel, for example, “K, K, K, For example, if a consonant first appears and the vowel intensity increases with the passage of time, or a complex combination of consonants and vowels, such as English, such as "ke, ko, sa, ta" However, the following problem occurs. Conventionally, voices are recorded instantaneously, and the sound spectrum is analyzed by integrating the sound spectrum of the entire band at regular intervals, making it difficult to determine at what point the consonant changed to a vowel As a result, the discrimination rate of voice recognition has been reduced. To solve this problem, more voice patterns are stored in the computer in advance and applied to any of these voice patterns, but this causes an increase in the computational load. ing.

The present invention has been made in view of such circumstances, and it is an object of the present invention to provide an acoustic sensor that can perform acoustic signal detection and frequency spectrum analysis on one piece of hardware at high speed and accurately.

Another object of the present invention is to provide an acoustic sensor which can perform acoustic signal detection, frequency spectrum analysis and frequency scale conversion (conversion to mel scale or bark scale) on one piece of hardware at high speed and accurately. To provide.

The acoustic sensor according to claim 1 includes a wave receiving portion that receives a sound wave propagating in a medium, a resonance portion having a plurality of rod-shaped resonators each having a length that resonates at a different specific frequency, It has a holding portion for holding a resonance portion, and a vibration intensity detection portion for detecting the vibration intensity of each of the resonators for each of the specific frequencies, and a distance between two adjacent resonators is different. It is characterized by the following.

The acoustic sensor according to claim 2, a receiving portion that receives a sound wave propagating in a medium, a resonant portion having a plurality of rod-shaped resonators each having a length such that each resonates at a different specific frequency, A holding portion for holding a resonance portion, and a vibration intensity detection portion for detecting the vibration intensity of each of the resonators for each of the specific frequencies, wherein a distance between two adjacent resonators is different. The bandwidth of the resonance frequency of each resonator is set to a predetermined value.

According to a third aspect of the present invention, in the acoustic sensor according to the first or second aspect, the resonance frequencies of the plurality of resonators are set to be distributed on a mel scale.

According to a fourth aspect of the present invention, in the acoustic sensor according to the first or second aspect, the resonance frequencies of the plurality of resonators are set to be distributed on a Bark scale.

According to a fifth aspect of the present invention, in the acoustic sensor according to the first or second aspect, the resonance frequencies of the plurality of resonators are set so as to be distributed on a bark scale, and a bandwidth corresponding to each resonance frequency is a critical bandwidth. It is characterized by being.

(6) The acoustic sensor according to claim 6 is the music song input microphone for recognizing a music song in claim 3.

(7) The acoustic sensor according to claim 7 is the audio input microphone for recognizing a voice according to any one of claims 1 to 5.

音響 The acoustic sensor according to claim 8 is characterized in that, in any one of claims 1 to 7, the acoustic sensor is formed on a semiconductor substrate.

The first acoustic sensor of the present invention has a plurality of resonators each having a different length so that each resonator resonates at a specific frequency, and transmits a sound wave propagated in a medium to these resonators. Detect vibration. Then, the detected vibration amplitude is converted into an electric signal, and the electric signal is input to the integrating means, and the input electric signal is integrated in an arbitrary period. Then, the result of the integration is output for each specific frequency at any period.

In addition, the second acoustic sensor of the present invention has the same configuration as the first acoustic sensor, but does not distribute the resonance frequency of each resonator on a mathematically linear scale but linearly on a mel scale. Make it distributed. Since the correspondence between the actual vibration frequency and the mel scale is determined based on the above equation (1) and FIG. 6, the design specifications of each resonator can be easily determined. Then, by detecting vibration in each resonator in accordance with the mel-scale specification, and thereafter performing the same processing as that of the above-described first acoustic sensor, a physical quantity corresponding to the spectrum of the acoustic signal can be detected on the mel-scale.

The third acoustic sensor of the present invention has the same configuration as the first acoustic sensor, but the resonance frequency of each resonator is not linearly distributed on a mathematically linear scale but linearly on a Bark scale. In addition to the distribution, the bandwidth of each resonance frequency is set to be a critical bandwidth. Since the correspondence between the actual vibration frequency and the Bark scale and the cutoff frequency that determines the critical bandwidth are determined based on the equations (2) and (3) and FIG. 7, the design specifications of each resonator can be easily determined. Can decide. Then, by detecting the vibration in each resonator in accordance with the Bark scale specification, by performing the same processing as the first acoustic sensor described above, the physical quantity corresponding to the spectrum of the acoustic signal with a critical bandwidth It can be detected on the Bark scale.

音響 The acoustic sensor of the present invention can detect the sound intensity for each desired frequency, so that an acoustic spectrum can be obtained in real time without performing analysis processing. Therefore, compared to the conventional method of inputting the audio signal of the entire band and electrically filtering each frequency band, in the present invention in which the acoustic signal is mechanically decomposed for each frequency, the electrical filtering is performed. It becomes unnecessary and the processing speed increases. Also, there is no loss of acoustic data anywhere even if the audio data is divided at regular intervals. In addition, since sound data for each frequency is obtained at regular time intervals, it is possible to check the transition of the intensity of each frequency as time passes, for example, it is possible to more accurately determine the temporal change between vowels and consonants In addition, the discrimination rate of voice recognition can be increased.

  As described above, in the acoustic sensor of the present invention, the sound wave is mechanically decomposed for each frequency band before being converted into an electric signal, so that there is no need for conventional electrical filtering processing using software. And the processing speed increases. Further, since the semiconductor device can be easily manufactured on a semiconductor substrate, an occupied area can be reduced as compared with a conventional system, and cost can be reduced. Furthermore, since the sound intensity can be detected for each desired frequency, an acoustic spectrum can be obtained in real time without performing an analysis process, and since sound data for each frequency can be obtained at regular time intervals, The transition of the intensity of each frequency can be confirmed as time elapses, and the temporal change of voice can be determined more accurately, which can contribute to increasing the recognition rate of voice recognition.

Further, the acoustic sensor of the present invention is an aggregate of a plurality of resonators having a resonance frequency distributed on a mel scale, or an aggregate of a plurality of resonators having a resonance frequency distributed on a bark scale and having a critical bandwidth. Therefore, the voice can be recognized in a state approximated by human hearing, and the characteristics of the voice can be efficiently extracted at the time of voice recognition.

Hereinafter, the present invention will be specifically described with reference to the drawings showing the embodiments.

(First Embodiment)
FIG. 1 is a diagram showing an embodiment of an acoustic sensor of the present invention. The acoustic sensor of the present invention includes a sensor main body 2 formed on a semiconductor silicon substrate 1, electrodes 3 and a detection circuit 4 as a peripheral circuit. The sensor main body 2 has a plurality of (six in the example of FIG. 1) rod-shaped portions having different lengths, all of which are formed of semiconductor silicon. A plate-shaped holding portion 22 held on the end side, a short rod-shaped propagation portion 23 erected at one end of the holding portion 22, and a plate shape connected to the propagation portion 23 and receiving a sound wave propagated in the air And a wave receiving portion 24 of FIG.

  The resonance portion 21 is a cantilever, and each bar-shaped portion is a resonator 25 whose length is adjusted to resonate at a specific frequency. The plurality of resonators 25 selectively vibrate at a resonance frequency f expressed by the following equation (4).

f = (CHE ^1/2 ) / (L ² ρ ^1/2 ) (4)
Where C: constant determined experimentally H: thickness of each resonator L: length of each resonator E: Young's modulus of material (semiconductor silicon) ρ: density of material (semiconductor silicon)

分 か る As can be seen from the above equation (4), by changing the thickness H or the length L of the resonator 25, the resonance frequency f can be set to a desired value. In the example shown in FIG. 1, the thickness H of all the resonators 25 is constant, and the length L is set so as to gradually increase from the left side to the right side. I have to. Specifically, it is possible to cope with high to low frequencies within a range of about 15 to / 20 kHz of the audible band from left to right.

共振 The bandwidth of the resonance frequency of each resonator 25 depends on the interaction with the adjacent resonator 25 in the process of transmitting vibration energy through the resonance portion 21. That is, the bandwidth is determined by the change rate of the resonance frequency of the adjacent resonator 25, the structural design value such as the distance to the adjacent resonator 25, the viscosity of the gas between the adjacent resonators 25, and the like. Although determined, in this example, the bandwidth of the resonance frequency of each resonator 25 is controlled by changing the distance between the adjacent resonators 25.

FIG. 5 is a graph showing a change in the bandwidth (vertical axis) of the single crystal silicon resonator 25 having a resonance frequency of 3 kHz when the distance D (horizontal axis) to the adjacent resonator 25 is changed. It is. FIG. 6 is a diagram showing the relationship among the length L, the thickness H, the width W, and the distance D in the resonator 25. The design values of the resonator 25 are: length L = 1706 μm, thickness H = 10 μm, width W = 80 μm, and the gas between the adjacent resonators 25 is air. It is understood from FIG. 5 that a desired bandwidth can be set by adjusting the distance D to the adjacent resonator 25. Therefore, in consideration of this, in this example, the distance D between the adjacent resonators 25 is determined so that the bandwidth of each resonator 25 becomes the critical bandwidth shown in FIG.

The sensor main body 2 having the above configuration is manufactured on the semiconductor silicon substrate 1 by using a semiconductor integrated circuit manufacturing technique or a micromachining technique. In such a configuration, when a sound wave is transmitted to the wave receiving portion 24, the plate-shaped wave receiving portion 24 vibrates, and the vibration indicating the sound wave propagates through the propagation portion 23 to the holding portion 22 and is held there. The rod-shaped resonators 25 of the resonating portion 21 are transmitted from left to right in FIG. 1 while sequentially resonating at respective specific frequencies.

An appropriate bias voltage V _bias is applied to the sensor main body 2, and is applied to the tip of each resonator 25 of the resonance portion 21 and the electrode 3 formed on the semiconductor silicon substrate 1 at a position facing the tip. Thus, a capacitor is formed. The tip of the resonator 25 is a movable electrode whose position moves up and down with the vibration of the resonator 25, while the electrode 3 formed on the semiconductor silicon substrate 1 is a fixed electrode whose position does not move. Then, when the resonator 25 vibrates at each specific frequency, the distance between the two electrodes changes, so that the capacitance of the capacitor changes.

Each electrode 3 is connected to a detection circuit 4 which converts such a change in capacitance into a voltage signal, integrates the converted voltage signal within a predetermined time, and outputs the integrated signal. Figure 2 is a diagram showing a configuration of the detection circuit 4, the detection circuit 4 includes an operational amplifier 41 for amplifying at amplification ratio corresponding to an impedance ratio between the capacitor C _s and the reference capacity C _f of the capacitor An integrating circuit 43 for integrating the output signal of the operational amplifier 42 higher than the reference voltage _Vref for a predetermined time, and a sample and hold circuit 44 for taking out the output signal from the integrating circuit 43, temporarily holding the output signal, and outputting it. The detection circuit 4 having such a configuration is formed by, for example, a silicon CMOS process.

Clock pulses φ ₀ , φ _1, and φ ₂ are supplied to the operational amplifier 41, the integrating circuit 43, and the sample-and-hold circuit 44, respectively. Work. Note that these clock pulses may be supplied from outside, or a counter circuit may be formed on the same semiconductor silicon substrate and supplied from there.

Next, the operation will be described. When the sound wave propagated in the air is transmitted to the wave receiving portion 24 of the sensor main body 2, the plate-like wave receiving portion 24 vibrates, and the vibration propagates in the sensor main body 2. At this time, the sound wave propagates from the left to the right in FIG. 1 while resonating the cantilevered resonators 25 whose length is gradually increased. Each resonator 25 has a unique resonance frequency, and each resonator 25 resonates when a sound wave of the unique frequency propagates, and its tip vibrates up and down. Due to this vibration, the capacitance of the capacitor formed between the tip and the electrode 3 changes. Since the energy of the sound wave is sequentially converted into the vibration energy of the resonator 25 as the sound wave propagates, the energy of the sound wave is gradually attenuated by such resonance, and the longest resonator 25 (see FIG. 1) By the time the sound wave reaches the right end), the energy as the sound wave has almost disappeared, and no reflected wave occurs. Therefore, there is no possibility that the reflected wave affects the capacitance change, and an accurate capacitance change matching the spectrum of the propagated sound wave can be detected.

The obtained capacitance change is sent to the detection circuit 4. FIG. 3 is a diagram showing a timing chart in the detection circuit 4, showing clock pulses φ ₀ , φ _1, and φ ₂ supplied to the operational amplifier 41, the integrating circuit 43, and the sample hold circuit 44, respectively. Note that the clock pulse control in this example is turned on at a low level.

First, within the detection circuit 4, the amplification ratio is determined in accordance with the impedance ratio between the capacitance C _s and the reference capacity C _f of the capacitor obtained by the operational amplifier 41. For example, when the value of 1 / ωC _s with respect to 1 / ωC _f (ω = 2πf, f: frequency) is １ ／, the obtained voltage signal is doubled. However, since the operational amplifier 41 is an inverting amplifier whose + input terminal is grounded, the voltage phase of the operational amplifier 42 at the next stage is inverted by one. The obtained amplified voltage signal is input to the integrating circuit 43. The integrating circuit 43, high amplification voltage signal from the reference voltage V _ref in a predetermined corresponding to the clock pulses phi ₁ time is accumulated, the integrated signal is input to the sample-and-hold circuit 44. The sample-and-hold circuit 44, and outputs the integrated signal to the outside by repeating the sampling and holding of the integrated signal in response to clock pulses phi _2.

The processing as described above is performed in parallel for each of the detection circuits 4 corresponding to the resonators 25 having different lengths. The cycle of the clock pulses φ ₀ , φ _1, and φ ₂ shown in FIG. 3 is merely an example, and the cycle of each clock pulse may be set arbitrarily.

As described above, according to the present invention, by examining the output signal of the detection circuit 4 corresponding to the resonator 25 that resonates at a specific frequency, the sound intensity of the specific frequency with an arbitrary time period is determined. Can be known over time. Further, by examining the output signals of the detection circuit 4 corresponding to the plurality of resonators 25, it is possible to know the temporal change of the sound intensity for each of a plurality of frequency bands with an arbitrary time period.

FIG. 4 is a diagram illustrating a relationship between the respective detection circuits 4 corresponding to a specific frequency. For example, n kinds of resonant frequencies _{f 1, f 2, f 3} , f 4, ..., in the case where each f _n selectively providing the n number of resonators to respond vibrations that each resonant frequency It is possible to obtain output signals V ₁ , V ₂ , V ₃ , V ₄ ,..., V _n of the detection circuit according to the resonance intensity. For example, when using the acoustic sensor of the present invention as a microphone for speech input for speech recognition, the strength of each resonance frequency in the audible band is determined according to the resonance strength, and based on the determined analysis pattern. To recognize the voice.

When it is desired to obtain the intensity of only the arbitrarily selected frequency of the sound wave, only the output signal of the detection circuit corresponding to the required resonance frequency may be obtained. For example, when the intensities of the frequencies f ₁ and f ₃ are obtained in FIG. 4, the outputs of the other detection circuits 4-2, 4-4,... By not providing 4-2, 4-4,..., 4-n, necessary output signals V ₁ , V ₃ are obtained, and unnecessary output signals V ₂ , V ₄ ,. _What is necessary is to make it impossible to obtain _n . As an example of use of such an acoustic sensor, an abnormal sound input microphone for detecting an abnormal sound of one or more specific frequencies is suitable.

(Second embodiment)
Next, a description will be given of a second embodiment in which the resonance frequency of each resonator is linearly distributed on a mel scale, which is a psychological attribute representing the pitch of a sound as represented by a musical scale. The configuration of the acoustic sensor according to the second embodiment is the same as that of the above-described first embodiment. However, in the second embodiment, the resonance frequency of each resonator 25 is mathematically determined. Instead of being distributed on a linear scale, the distribution is made linear on the mel scale. That is, when the resonance frequencies of the n resonators 25 are f ₁ , f ₂ , f ₃ ,..., F _n ,
f ₁ [Hz] = αf ₂ [Hz] =... = α ^n-1 f _n [Hz]
Instead of setting like
f ₁ [mel] = αf ₂ [mel] =... = α ^n-1 f _n [mel]
Set as follows. Here, α is a coefficient that can be arbitrarily set.

The resonance frequency of each resonator 25 is determined by the above equation (4), and the correspondence between the actual vibration frequency and the mel scale is determined based on the above equation (1) and FIG. Therefore, an arbitrary resonance frequency on the mel scale can be easily assigned to each resonator 25. In this example, the thickness H of all the resonators 25 is constant, and the length L is varied to obtain a resonance frequency corresponding to a frequency that is equally spaced on the mel scale.

Note that other configurations and operations are the same as those in the first embodiment, and thus description thereof is omitted.

In the second embodiment, the resonance frequencies of the respective resonators 25 are distributed on a mel scale, so that octave sounds, semitones, etc., which can be heard by human ears can be selectively recognized in real time. It is possible to produce a microphone having a matched frequency characteristic. Since the temporal change of pitch sounds such as octave sounds and semitones can be more accurately discriminated, it is effective not only for voice recognition and abnormal sound detection, but also for musical scales such as reading, reading with inflections such as waka, and music. It is possible to configure a microphone for voice input that is excellent in sound discrimination.

(Third embodiment)
Next, a description will be given of a third embodiment in which the resonance frequency of each resonator is linearly distributed on a Bark scale, which is a psychological attribute representing the loudness of a sound. The configuration of the acoustic sensor according to the third embodiment is similar to the configuration of the above-described first embodiment. However, in the third embodiment, the resonance frequency of each resonator 25 is mathematically determined. Instead of being distributed on a linear scale, the distribution is made on a Bark scale, and the bandwidth of the resonance frequency in each resonator 25 is made to be a critical bandwidth.

共振 The resonance frequency of each resonator 25 is determined based on the correspondence between the bark scale and the actual frequency shown in FIG. The resonance frequency of each resonator 25 is determined by the above equation (4). In this example, the thickness H of all the resonators 25 is fixed, and the length L is made different to obtain the Bark scale. Are assigned to the respective resonators 25.

In the third embodiment, the resonance frequency of each resonator 25 is distributed on a bark scale, so that it is possible to have a frequency characteristic and a bandwidth suitable for human hearing and to hide the noise in noise. It becomes easy to select the sound signal that is present, and it is possible to improve the discrimination rate of speech recognition in a noisy situation. In addition, a sensor closer to human hearing can be provided.

Note that other configurations and operations are the same as those in the first embodiment, and thus description thereof is omitted.

It is a figure showing an embodiment of an acoustic sensor of the present invention. It is a figure showing composition of a detection circuit in an acoustic sensor of the present invention. It is a figure showing a timing chart of a detection circuit in an acoustic sensor of the present invention. FIG. 3 is a diagram illustrating a relationship between respective detection circuits corresponding to a specific frequency. 5 is a graph illustrating a relationship between a resonator distance and a bandwidth. It is a figure showing the relationship of the length, thickness, width, and distance of the resonator in the acoustic sensor of the present invention. 6 is a graph showing a relationship between an actual frequency and a mel scale value. 5 is a chart showing a numerical relationship between a critical bandwidth and a Bark scale.

Explanation of reference numerals

DESCRIPTION OF SYMBOLS 1 Semiconductor silicon substrate 2 Sensor body 3 Electrode 4 Detection circuit 21 Resonance part 22 Retention part 23 Propagation part 24 Receiving part 25 Resonator 41, 42 Operational amplifier 43 Integration circuit 44 Sample hold circuit

Claims (8)

  A receiving part that receives a sound wave propagating in a medium, a resonance part having a plurality of rod-shaped resonators each having a length such that each resonates at a different specific frequency, and a holding part that holds the resonance part, An acoustic sensor, comprising: a vibration intensity detection portion for detecting the vibration intensity of each of the resonators at each of the specific frequencies, wherein a distance between two adjacent resonators is different.   A receiving part that receives a sound wave propagating in a medium, a resonance part having a plurality of rod-shaped resonators each having a length such that each resonates at a different specific frequency, and a holding part that holds the resonance part, A vibration intensity detection portion for detecting the vibration intensity for each specific frequency of each of the resonators, and changing the distance between two adjacent ones of the resonators so that a resonance frequency band in each of the resonators is provided. An acoustic sensor having a width set to a predetermined value.   The acoustic sensor according to claim 1, wherein resonance frequencies of the plurality of resonators are set to be distributed on a mel scale.   The acoustic sensor according to claim 1, wherein resonance frequencies of the plurality of resonators are set so as to be distributed on a bark scale.   The acoustic wave according to claim 1, wherein the resonance frequencies of the plurality of resonators are set so as to be distributed on a bark scale, and a bandwidth corresponding to each resonance frequency is a critical bandwidth. Sensors.   The acoustic sensor according to claim 3, wherein the acoustic sensor is a music song input microphone for recognizing a music song.   The acoustic sensor according to claim 1, wherein the acoustic sensor is a voice input microphone for recognizing voice.   The acoustic sensor according to claim 1, wherein the acoustic sensor is formed on a semiconductor substrate.

Images